Skip to content

Latest commit

 

History

History
11 lines (10 loc) · 935 Bytes

File metadata and controls

11 lines (10 loc) · 935 Bytes

Customer_Tech_NonTech_Support_AgenticAI_CrewAI

Streamlit based Chatbot for Tech and Non-Tech Customer Queries using CrewAI framework to showcase RAG based AgenticAI in python v3.11.0

  • This Repo includes sample inputs/PDFs used, requirements.txt, screenshot of the chatbot, scripts
  • Script uses Azure OpenAI embeddings and LLM(gpt4o)
  • Install packages as "python.exe -m pip install -r Py3p11p0_CustSupport_CrewAI_requirements.txt"
  • I'm using unstructured.io for parsing the Tech PDF and Images(process is shown in UnstructuredIO_Parse_TechData.py), and Non-Tech PDF using CrewAI Tools
  • unstructured.io requires tesseract-ocr and poppler, extract, install and mention the installed path in the script
  • Storing parsed data in ChromaDB and using Langchain's MultiVectorRetriever
  • Modify the paths as required in the scripts
  • Run chatbot script as "python.exe -m streamlit run Customer_Tech_NonTech_Support_AgenticRAG_CrewAI.py"